Reliability and availability analysis of repairable systems often relies on commercial simulation tools that are costly, closed-source, and restricted to a fixed catalogue of failure and repair distributions. This paper presents farofa, an open-source Python framework for building discrete-event simulation models of repairable equipment fleets. The framework is modular: users compose models by combining devices with either built-in or user-defined statistical distributions for failure and repair behavior, along with configurable maintenance team and fleet parameters. This design supports a range of applications, including maintenance team sizing, reproducing and extending published reliability models, experimenting with non-standard distributions, and coupling simulation with optimization routines for parameter tuning. We describe the framework's architecture and main components, and illustrate its use through application examples. The project is released under the GNU GPL v3 and is publicly available on GitHub, with the goal of providing a transparent, extensible foundation for both research, practice and education in repairable-system reliability modeling.